Methodological Considerations in Scale Refinement with Diverse Populations: A Case Example Using the CES-D with a Community Sample of American Indian Women

Abstract

Although many psychometric assessments are used extensively in population-based research to determine psychopathology, these tools have not been thoroughly validated or appropriately adapted for use in diverse populations. Indeed, depression measurement studies among American Indian and female populations are scarce, omitting key opportunities to tailor psychological measurement for this population. To build psychometric evidence of measures in this population, we used a procedural method to examine a standard psychological instrument—the Center for Epidemiological Studies Depression Scale (CES-D)—with a community sample of southeast American Indian women. Our results showed strong psychometric reliability of the 20-item CES-D. The "effort" item presented diminished validity, as demonstrated by a negative counter-intuitive item-to-total correlation (ITC) value. Dropping the "effort" item resulted in a 19-item scale with a better fit in the within-group examination of community-based American Indian women. Compared to the 20-item CES-D scale, the revised 19-item measure ("effort" item removed) resulted in minimal changes to women’s depression categories. However, we did detect patterns in shifts such that the 19-item scale generally underestimated (i.e., placed women in a lower category) depressive symptoms compared to the 20-item scale. Depending on their study goals, researchers engaging in population-based research should carefully weigh the use of original scales that allow for consistency in reporting with refined scales that fit psychometrically. We present the outlined method as a tool that expands on current approaches in scale refinement, and aids researchers in making more informed decisions regarding refined scales with diverse populations.

Acknowledgments

We acknowledge the editorial assistance of Dr. Beth Black, which was supported by the University of North Carolina at Chapel Hill School of Nursing. We also recognize the valuable contribution of the Community Partnership Committee.

Disclosure statement

No potential conflict of interest was reported by the authors.

Data availability statement

The data supporting this study’s findings are available from the corresponding author, Dr. Jada Brooks, upon reasonable request.

Additional information

Funding

This work was supported by the National Institute of Environmental Health Science (NIEHS), National Institutes of Health (NIH), under Grant Award Number K23ES027026; National Heart, Lung, and Blood Institute (NHLBI), National Institutes of Health, under Grant Award Number K24HL105493; National Center for Advancing Translational Sciences (NCATS), National Institutes of Health, under Grant Award Number UL1TR002489; and U.S. Environmental Protection Agency (EPA) Assistance Agreement 83578501. The content is solely the responsibility of the authors and does not necessarily represent the official views of the NIH or EPA.